Fear and Free Speech

INTRODUCTION

Fear changes lives. And for this reason, fear sometimes changes the law.

Because of fear’s debilitating effects, the law often forbids behavior that causes its targets to fear for their physical safety. Think of the laws that prohibit threats in order to free their targets “to go about their own lives.”¹Kenneth L. Karst, Threats and Meanings: How the Facts Govern First Amendment Doctrine, 58 Stan. L. Rev. 1337, 1345 (2006). Think too of stalking laws that restrict fear-engendering behavior that “violates basic social norms of privacy and consent by persisting even in the face of a lack of consent.”²Genevieve Lakier & Evelyn Douek, The First Amendment Problem of Stalking: Counterman, Stevens, and the Limits of History and Tradition, 113 Calif. L. Rev. 143, 190 (2025).

A different type of fear—speakers’ fear of the government’s punishment—shapes pivotal First Amendment doctrine. The U.S. Supreme Court focused on this fear in Counterman v. Colorado when it held that the First Amendment requires the prosecutor in a “true threats” case to establish the speaker’s recklessness—that is, that the speaker “consciously disregarded a substantial risk” that his statements would make his target fear for her physical safety.³Counterman v. Colorado, 143 S. Ct. 2106, 2111–12 (2023). In the interest of full disclosure, I note that I served on the team representing the state of Colorado in Counterman before the U.S. Supreme Court. In so holding, the majority underscored that requiring such proof “reduc[es] an honest speaker’s fear that he may accidentally [or erroneously] incur liability,” thus “provid[ing] ‘breathing room’ for more valuable speech.”⁴Id. at 2115 (quoting United States v. Alvarez, 567 U.S. 709, 733 (2012) (Breyer, J., concurring)).

But this doctrinal choice is not without its costs, costs borne by the many targets of threats who lack access to evidence of the speaker’s interior mental state. This is the case, for example, of “[a] delusional speaker [who] may lack awareness of the threatening nature of her speech; a devious speaker [who] may strategically disclaim such awareness; and a lucky speaker [who] may leave behind no evidence of mental state for the government to use against her.”⁵Counterman, 143 S. Ct. at 2141 (Barrett, J., dissenting); see also Farmer v. Brennan, 511 U.S. 825, 836–37 (1994) (explaining that criminal recklessness requires a subjective showing that the defendant disregarded a risk of harm of which they were aware).

Free speech, of course, is not always free for everyone. A great deal of First Amendment law requires the targets of harmful speech to pay the price for the speaker’s freedom to speak and for the public’s freedom to receive that expression. To illustrate, the Supreme Court’s defamation jurisprudence sometimes requires innocent targets to bear the costs of reputation-damaging falsehoods when those targets cannot prove the speaker’s recklessness.⁶Gertz v. Robert Welch, Inc., 418 U.S. 323, 342–43 (1974). So too did a grieving father pay, with his pain, for the protected expression of speakers who exploited his loss to amplify their religious and political views.⁷Snyder v. Phelps, 562 U.S. 443, 460–61 (2011). Doctrinal choices are sometimes difficult because they require courts to pick between important, but incommensurable, values. When push comes to shove, First Amendment law generally chooses to protect speech on matters of public concern—the speech that lies at the core of the First Amendment—at the expense of its targets’ reputation and at the expense of its targets’ freedom from emotional cruelty.⁸E.g., N.Y. Times Co. v. Sullivan, 376 U.S. 254, 270 (1964) (emphasizing “the principle that debate on public issues should be uninhibited, robust, and wide-open, and that it may well include vehement, caustic, and sometimes unpleasantly sharp attacks on government and public officials”).

But different, and difficult, doctrinal choices arise when we recognize that speech interests sometimes lie on both sides of a First Amendment dispute, thus requiring courts to choose between important and commensurate values. As we will see, the fear induced by threats and by stalking can silence targets’ expression at least as directly and frequently as the fear induced by the prospect of the government’s punishment silences speakers’ expression. Consider the many targets who stop going out in public, stop using the phone, stop engaging online, stop making music and other forms of art, and stop participating in public life.⁹See infra notes 62–71 and accompanying text. Identifying free speech as a “preferred” constitutional value should thus require attention to the free speech costs of law’s failure to adequately protect the targets of threats and stalking.¹⁰See Frederick Schauer, Fear, Risk and the First Amendment: Unraveling the “Chilling Effect”, 58 B.U. L. Rev. 685, 732 (1978) (describing the First Amendment as a “preferred value”).

Yet an ostensibly speech-protective Supreme Court failed to acknowledge those costs in Counterman v. Colorado, a case that required it to define the contours of the category of threats unprotected by the First Amendment. “True threats” cause their targets to fear for their physical safety—and this fear, in turn, disrupts those targets’ lives.¹¹See R.A.V. v. City of St. Paul, 505 U.S. 377, 388 (1992) (explaining that the law punishes threats to protect their targets “from the fear of violence, from the disruption that fear engenders, and from the possibility that the threatened violence will occur”). For these reasons, the Supreme Court has long treated true threats as among the handful of speech categories entirely unprotected by the First Amendment.¹²Id. Until Counterman, however, the Supreme Court had never directly engaged the question of how to determine when speech constitutes an unprotected threat. In the meantime, lower courts disagreed over whether the universe of unprotected threats should be determined through a subjective, or instead, an objective inquiry. Some took a subjective approach, requiring the prosecution (or, in a civil case, the plaintiff) to prove that the speaker intended to make his target fear for her physical safety, thus privileging defendants’ speech interests.¹³See, e.g., United States v. Bachmeier, 8 F.4th 1059, 1064 (9th Cir. 2021); United States v. Heineman, 767 F.3d 970, 975 (10th Cir. 2014). A majority of lower courts instead approached the problem from the target’s perspective, adopting an objective approach that asked whether the speaker’s statement would cause a reasonable person in the target’s position to fear for their physical safety, thus privileging targets’ interests—including, but not limited to, their speech interests.¹⁴See, e.g., United States v. Nishnianidze, 342 F.3d 6, 15 (1st Cir. 2003); Heller v. Bedford Cent. Sch. Dist., 665 F. App’x 49, 51 n.1 (2d Cir. 2016); United States v. White, 670 F.3d 498, 509 (4th Cir. 2012); Porter v. Ascension Par. Sch. Bd., 393 F.3d 608, 616 (5th Cir. 2004); United States v. Stewart, 411 F.3d 825, 828 (7th Cir. 2005); United States v. Ivers, 967 F.3d 709, 718 (8th Cir. 2020).

The Supreme Court resolved this question in Counterman by requiring a subjective standard, holding that speech communicating a serious intent to commit unlawful violence constitutes an unprotected “true threat” only when the prosecution can prove the defendant’s recklessness—in other words, the defendant’s conscious disregard of the risk that his statements could make his target fear for her physical safety.¹⁵Counterman v. Colorado, 143 S. Ct. 2106, 2111 (2023). The Counterman Court justified this choice as speech-protective, emphasizing that speakers’ expression might be chilled if they fear punishment under a legal standard too quick to view their speech as threatening.¹⁶Id. at 2113 (“Counterman contends . . . that the absence of such a mens rea requirement will chill protected, non-threatening speech. . . . To combat the kind of chill he references, our decisions have often insisted on protecting even some historically unprotected speech through the adoption of a subjective mental-state element. We follow the same path today, holding that the State must prove in true-threats cases that the defendant had some understanding of his statements’ threatening character.”).

But this was not the only speech-protective choice available to the Court. In choosing to privilege the defendant’s speech over the target’s reasonable fear for their physical safety, the majority also effectively chose to privilege the defendant’s speech over the target’s expression. And just as the public loses valuable speech when the prospect of governmental punishment deters speakers from speaking, so too does the public lose valuable speech when the law permits the silencing of targets’ speech.¹⁷See infra notes 30–31, 62–71 and accompanying text.

To be sure, we should worry about doctrinal choices that deter valuable speech by making a potential speaker fear that the government will punish them for what they intended as a joke or as political rhetoric. Recall, for example, eighteen-year-old Robert Watts, who was convicted of threatening President Lyndon B. Johnson during the Vietnam War for saying to laughing listeners at a political rally on the Washington, D.C. Mall that “I have already received my draft classification as 1–A and I have got to report for my physical this Monday coming. I am not going. If they ever make me carry a rifle the first man I want to get in my sights is L. B. J.”¹⁸Watts v. United States, 394 U.S. 705, 706 (1969) (per curiam). The Supreme Court overturned Watts’s conviction, holding that his speech constituted “political hyperbole” rather than a “true threat” against the president.¹⁹Id. at 708 (internal quotation marks omitted).

At the same time, however, we should also worry about doctrinal choices that enable the silencing of targets’ expression by insufficiently protecting those targets from reasonable fear for their physical safety. Think now of musician C.W., the victim in Counterman, who received hundreds of disturbing messages from a complete stranger who repeatedly resisted her efforts to block him. As a result, C.W. stopped walking alone, stopped attending public events, and canceled her own public musical performances.²⁰See Counterman, 143 S. Ct. at 2112. Think too of the many threats to journalists and election officials that silence and suppress their work foundational to a free and fair democracy.²¹See infra notes 65–71 and accompanying text. The speech-silencing effects of fear-inducing speech are by no means limited to artists and journalists: they extend to all targets, because all are potential speakers.

Courts often resolve difficult First Amendment disputes—including that in Counterman, which called upon the Court to define the contours of a category of unprotected speech—by concluding that the defendant’s speech interests outweigh the target’s substantial but incommensurate interests in reputation or peace, among other values. But threats and stalking cases present even greater challenges for those who consider speech to be a preferred value because commensurate speech interests lie on both sides of these disputes.

The First Amendment issues triggered by the regulation of fear-inducing speech thus require hard choices between speakers’ and targets’ free speech interests, choices that also affect the public’s access to valuable speech. This Article both criticizes the Court’s approach in Counterman and offers a guide to addressing targets’ expressive interests in threats and stalking cases after Counterman. More specifically, it seeks to foster a mindset for considering the First Amendment problems involving fear-inducing speech that attends to targets’ expression as much as defendants’ expression. To this end, it examines the ways in which threats and stalking can silence targets’ speech and deter their participation in public life, and then considers available choices that attend to targets’—not just defendants’—expressive interests.

I. Fear on Both Sides, Speech on Both Sides

This Part starts by describing concerns that fear of government punishment will cause speakers to censor themselves, thus “chilling” their speech. It then explains how the fear induced by threats and stalking can cause their targets to engage in self-censorship of their own.

A. Speakers’ Fear of the Government

 “A chilling effect,” Frederick Schauer explained, occurs “when individuals seeking to engage in activity protected by the first amendment are deterred from so doing by governmental regulation not specifically directed at that protected activity.”²²See Schauer, supra note 10, at 693 (emphasis omitted). In other words, concerns about chilling effects anticipate that government’s efforts to modify some unprotected behavior will inadvertently chill protected behavior. As Leslie Kendrick observed, “intuition suggests that some legal rules will chill speech. The further a law encroaches on protected speech, the greater the risk that such speech will be penalized. The more likely speakers are to be penalized, the less they will speak.”²³Leslie Kendrick, Speech, Intent, and the Chilling Effect, 54 Wm. & Mary L. Rev. 1633, 1638 (2013).

Because chilling effects harm speakers—and the rest of us, too—by stifling the delivery of opinions, facts, and ideas, the Court’s instrumental concerns about chilling effects not infrequently drive its doctrinal choices.²⁴See Monica Youn, The Chilling Effect and the Problem of Private Action, 66 Vand. L. Rev. 1473, 1483 (2013) (“[T]he chilling effect concept does not delineate a discrete, freestanding doctrinal category. . . . [Instead, it informs] a number of procedural, categorical, and substantive doctrines in First Amendment case law.”). Under its defamation jurisprudence, for example, the more valuable the speech potentially chilled by the prospect of defamation liability, the higher the bar for proving the mental state required to establish such liability.²⁵See N.Y.  Times Co. v. Sullivan, 376 U.S. 254, 271–72, 279–80 (1964). This doctrine thus requires defamation plaintiffs who are public officials or public figures to show that a speaker made a defamatory statement with “ ‘actual malice’—that is, with knowledge that it was false or with reckless disregard of whether it was false or not”—before they can establish the speaker’s liability for reputation-damaging falsehoods.²⁶Id. at 279–80. The Court has explained this doctrinal innovation as necessary to ensure breathing space for speech critical of those in the public eye or otherwise on matters of public concern—the speech at the core of the First Amendment.²⁷Id. at 271–72.

In the same vein, the Counterman Court explained its doctrinal choice as “based on fear of ‘self-censorship’—the worry that without such a subjective mental-state requirement, the uncertainties and expense of litigation will deter speakers from making even truthful statements.”²⁸Counterman v. Colorado, 143 S. Ct 2106, 2115 (2023) (quoting N.Y. Times, 376 U.S. at 279). In explaining its decision to require the prosecution in a “true threats” case to show the defendant’s reckless mental state, the Court catalogued the ways in which the prospect of the government’s punishment can chill a speaker’s expression by instilling fear of various consequences: “The speaker’s fear of mistaking whether a statement is a threat; his fear of the legal system getting that judgment wrong; his fear, in any event, of incurring legal costs—all those may lead him to swallow words that are in fact not true threats.”²⁹Counterman, 143 S. Ct at 2116; see also id. at 2115 (“The result is ‘self-censorship’ of speech that could not be proscribed—a ‘cautious and restrictive exercise’ of First Amendment freedoms.”) (quoting Gertz v. Welch, 418 U.S. 323, 340 (1974)).

True, so far as it goes. But nowhere did the Counterman majority similarly detail targets’ fear for their physical safety and the resulting costs of such fear—costs that include the silencing of their speech along with many other life disruptions. The silencing of targets’ speech harms the general public, too, by denying it the expression that targets would otherwise deliver.³⁰See infra notes 64–71 and accompanying text. The majority instead simply noted the “profound harms” imposed by threats without specifying those harms.³¹Counterman, 143 S. Ct at 2117 (noting “the profound harms, to both individuals and society, that attend true threats of violence—as evidenced in this case”). Justice Sotomayor’s concurring opinion was more direct, though brief, in acknowledging the harms to targets: “Stalking can be devastating and dangerous. Lives can be ruined, and in the most tragic instances, lives are lost. . . . Even isolated threatening speech can do real harm. Such speech not only disrupts lives, it can silence the speech of others who become afraid to speak out.” Id. at 2123 (Sotomayor, J., concurring in part and concurring in the judgment) (citation omitted).

Judges (like other human beings) sometimes mask the difficulty of hard trade-offs—making those trade-offs feel considerably easier than they actually are—by failing to identify and weigh the costs of their choices. As Mary Anne Franks observed, courts’ “concern about chilling effects tends to be highly selective.”³²Mary Anne Franks, Fearless Speech, 17 First Amend. L. Rev. 294, 306 (2019). Others have also observed this dynamic. See Danielle Keats Citron, From Bad to Worse: Stalking, Threats, and Chilling Effects, 2023 Sup. Ct. Rev. 175, 180 (2023); R. George Wright, Counterman v. Colorado: True Threats, Speech Harms, and Missed Opportunities, 99 Ind. L.J. 27, 30 (2023). Indeed, even as the Counterman majority described

itself as balancing competing interests, it particularized only speakers’ interests.³³See Counterman, 143 S. Ct at 2119.

The majority’s choice may have been motivated not only by its instrumental concerns about chilling defendants’ expression but also by deontic commitments to punish only morally blameworthy speakers. In this vein, the majority described “reckless defendants” as “morally culpable” because they “have done more than make a bad mistake. They have consciously accepted a substantial risk of inflicting serious harm.”³⁴Id. at 2118; see also Kendrick, supra note 23, at 1633 (suggesting that courts’ intuitions about moral culpability contribute to their doctrinal choices in this area because “the difficulties of measuring and remedying chilling effects cast doubt on whether they could ever provide the sole justification for the choice of one intent requirement over another”). But moral principles cut both ways here, too, as the targets of threats themselves are often entirely innocent victims who experience life-changing harm regardless of the speaker’s mental state. Think again of C.W., the victim in Counterman whose daily existence and musical career were “upended” by hundreds of disturbing messages from a man she had never met.³⁵Counterman, 143 S. Ct at 2112. Think too of journalists and election officials targeted by threats who live in fear for their, and their families’, safety for simply doing their jobs—jobs foundational to a functioning democracy.³⁶See infra notes 65–71 and accompanying text. That we can identify deontic as well as instrumental concerns on both sides of these disputes adds to the difficulties in deciding them, difficulties that judges should not obscure.³⁷See Seana Valentine Shiffrin, The Moral Neglect of Negligence, in 3 Oxford Studies in Political Philosophy 213 (David Sobel, Peter Vallentyne & Steven Wall eds., 2017) (describing the moral failures of negligence as involving “a failure to take and exercise appropriate responsibility for one’s agency; and, when that failure involves other people, negligence involves a failure properly to recognize and acknowledge their moral significance”).

B. Targets’ Fear for Their Physical Safety

Threats and stalking are related but distinct in important ways. I use the term “threat” to mean speech that causes its target to fear for their physical safety.³⁸See R.A.V. v. City of St. Paul, 505 U.S. 377, 388 (1992) (explaining that law punishes threats to protect their targets “from the fear of violence, from the disruption that fear engenders, and from the possibility that the threatened violence will occur”). And by “stalking,” I mean repeated and unwelcome conduct, contact, or communication that causes its target fear or other severe emotional distress.³⁹See, e.g., Colo. Rev. Stat. § 18-3-602(1)(c) (2023) (making it unlawful “[r]epeatedly [to] follow[], approach[], contact[], place[] under surveillance, or make[] any form of communication with another person . . . in a manner that would cause a reasonable person to suffer serious emotional distress and does cause that person . . . to suffer serious emotional distress”). Both threats and stalking can inflict life-disrupting fear. But they often do so in different ways: all threats, by definition, involve speech, while stalking sometimes involves only conduct, sometimes only speech, and sometimes a mix of the two.⁴⁰See infra notes 116–18 and accompanying text; see also Lakier & Douek, supra note 2 (“It is this violation of social norms and its disregard for the autonomy of the listener—her right to withdraw from the relationship and to not have to hear what the speaker has to tell her—that makes stalking speech so frightening. It also threatens the target’s sense of agency and freedom because of its (seemingly unstoppable) intrusion into the victim’s private life.”) And although a speaker may be held liable for a single threatening statement or for many, stalking laws impose liability only for “repeated” unwelcome conduct, contact, or communication.⁴¹See infra notes 114–15 and accompanying text.

Fear for one’s physical safety literally changes lives—and not for the better, as fear is among the most debilitating of human experiences. Consider psychologist Abraham Maslow’s hierarchy of human needs.⁴²A. H. Maslow, A Theory of Human Motivation, 50 Psych. Rev. 370, 394–95 (1943) (identifying physical safety as among the most important of human goals and thus among those that “monopolize” our time, energy, and attention). Foundational to this hierarchy is our need for physical survival in the form of food, water, shelter, and physical security: only when this need is satisfied can we pursue higher-level social and emotional needs like intimacy, belonging, self-esteem, confidence, and more.⁴³Id. For these reasons, requiring a target to pay for a defendant’s free speech with fear for their physical safety exacts a significantly higher price than requiring them to pay for a defendant’s free speech with their reputation or with their freedom from emotional cruelty.

Fear-inducing speech inflicts harm both physical and mental. One study, for example, reported “high levels of psychopathology among stalking victims,” whose overall psychological distress was about three times higher than that of the general population.⁴⁴Eric Blaauw, Frans W. Winkel, Ella Arensman, Lorraine Sheridan & Adriënne Freeve, The Toll of Stalking: The Relationship Between Features of Stalking and Psychopathology of Victims, 17 J. Interpersonal Violence 50, 57–58 (2002). Medical evidence demonstrates the life-changing effects of chronic fear that include insomnia and sleep disruptions,⁴⁵Id.; Jaime Rosenberg, The Effects of Chronic Fear on a Person’s Health, Neuroscience Educ. Inst. Cong. (Nov. 11, 2017), https://www.ajmc.com/view/the-effects-of-chronic-fear-on-a-persons-health [https://perma.cc/M9RM-7JSV]. immune system dysfunction,⁴⁶Rosenberg, supra note 45. gastrointestinal issues like ulcers and irritable bowel syndrome,⁴⁷Joe Pierre, How Does Fear Influence Risk Assessment and Decision-Making?, Psych. Today (July 15, 2020), https://www.psychologytoday.com/us/blog/psych-unseen/202007/how-does-fear-influence-risk-assessment-and-decision-making. sexual dysfunction and decreased fertility,⁴⁸Id. and accelerated aging.⁴⁹Mai Stafford, Tarani Chandola & Michael Marmot, Association Between Fear of Crime and Mental Health and Physical Functioning, 97 Am. J. Pub. Health 2076, 2078 (2007) (finding that some study participants had “limitations in physical functioning . . . that were commensurate with that of people 9 years” older). Medical evidence also establishes that chronic fear is associated with generally poor mental health,⁵⁰See Jessica Miles, Straight Outta SCOTUS: Domestic Violence, True Threats, and Free Speech, 74 U. Mia. L. Rev. 711, 735 (2020) (“In addition to an increased risk of physical violence, intimate partner stalking victims suffer from high rates of anxiety and depression.”). with mental health injuries that include anxiety,⁵¹Blaauw et al., supra note 44, at 57. depression,⁵²Id. post-traumatic stress disorders,⁵³Sean Wake, Jolie Wormwood & Ajay B. Satpute, The Influence of Fear on Risk Taking: A Meta-Analysis, 34 Cognition & Emotion 1143, 1148 (2020). panic and obsessive compulsive disorders,⁵⁴Id. impaired formation of long-term memories,⁵⁵Rosenberg, supra note 45. and suicidal ideation.⁵⁶Blaauw et al., supra note 44, at 57 (approximately one third of stalking victims reported repeated thoughts about committing suicide). One study found that those who reported higher levels of fear were 50% more likely to show signs of a mental health disorder and 90% more likely to have symptoms of depression than those who reported lower levels of fear.⁵⁷Stafford et al., supra note 49.

Moreover, the targets of threats and stalking often fear for their physical safety for very good reason. Criminal justice scholar Mary Brewster, for instance, canvassed the experiences of victims of intimate partner violence to conclude that threats served as “a strong and statistically significant predictor of violence.”⁵⁸Mary P. Brewster, Stalking by Former Intimates: Verbal Threats and Other Predictors of Physical Violence, 15 Violence & Victims 41, 50 (2000); see also Miles, supra note 50, at 734–35 (“Research on stalking demonstrates that a strong correlation exists between intimate partner stalking and physical violence, as eighty-one percent of women who were stalked by a current or former intimate partner were also physically assaulted by that partner.”). By causing targets to fear for their physical safety, threats and stalking not only inflict physical and mental health injuries, but also change the ways in which those targets live their lives.⁵⁹See Paul E. Mullen & Michele Pathé, Stalking, 29 Crime & Just. 273, 296 (2002) (“All but six of the one hundred victims [surveyed] reported major lifestyle changes and modified their daily activities in direct response to being stalked.”). Stalking, for instance, often forces its targets to miss work,⁶⁰See Patricia Tjaden & Nancy Thoennes, Nat’l Inst. of Just. & Ctrs. for Disease Control & Prevention, Stalking in America: Findings from the National Violence Against Women Survey 11 (1998), https://www.ojp.gov/pdffiles/169592.pdf [https://perma.cc/BLY2-25T5]; Mullen & Pathé, supra note 59, at 297. to leave their jobs, or their homes.⁶¹See Katrina Baum, Shannan Catalano, Michael Rand & Kristina Rose, Bureau of Just. Stat. Special Rep., Stalking Victimization in the United States 6 (2009) (“One in 7 victims reported they moved as a result of the stalking.”); Mullen & Pathé, supra note 59, at 296–97.

Most relevant to this Article, threats and stalking also change their targets’ lives by causing them to speak less, to speak differently, and to participate less in public life. Indeed, the targets of threats and stalking commonly, and reasonably, respond to threats to their physical survival by making themselves silent and invisible.⁶²See Jonathon W. Penney, Understanding Chilling Effects, 106 Minn. L. Rev. 1451, 1510–12 (2022) (“[T]hreats of violence and physical harm are a powerful force for self-censorship, which trigger deeper psychological states of fear, anxiety, and severe emotional distress that then in turn amplify social conformity.” (footnote omitted)). To illustrate, recall the musician’s experience in Counterman: C.W. stopped walking alone, stopped attending other musicians’ events, and canceled her own performances.⁶³See Counterman v. Colorado, 143 S. Ct. 2106, 2112 (2023). Danielle Keats Citron explains how stalking silences protected expression more generally:

[Cyber stalking victims] stop using their phones. Victims change how they express themselves; they are less controversial, more muted, and connect with fewer people. Their withdrawal from online engagement isolates them from friends and family. When victims change their phone numbers to prevent stalkers from calling them, they become unreachable.⁶⁴Citron, supra note 32, at 198–99.

Consider too the many journalists targeted by threats of death or rape. One recent study found that one in ten journalists surveyed had been threatened with death in the preceding year because of their work;⁶⁵Erin C. Carroll, Obstruction of Journalism, 99 Denv. L. Rev. 407, 415 (2022). another found that three quarters of female journalists surveyed reported that “they had experienced online abuse, harassment, threats, and attacks.”⁶⁶Id. at 409. These threats interfere with journalists’ watchdog and educator functions,⁶⁷See Vincent Blasi, The Checking Value in First Amendment Theory, 2 Am. Bar Found. Rsch. J. 521, 538–39 (1977) (explaining the press’s role as a watchdog for government misconduct); RonNell Andersen Jones, Press Speakers and the First Amendment Rights of Listeners, 90 U. Colo. L. Rev. 499, 537–43 (2019) (explaining the press’s functions as educating the public on a wide range of matters and serving as the public’s proxy by observing what the public does not have the time or resources or other ability to observe for itself). inflicting injury not only to their individual targets, but also to a public’s hopes for a healthy democracy. Free press scholar Erin Carroll has detailed more specifically how threats inhibit and distort the reporting that journalists produce and that the public thus receives:

Obstruction operates on at least three levels: the story, the beat, and the pipeline. At the story level, threats and abuse prevent journalists from covering particular events or incidents. At the beat level, perpetual abuse around broad topics like politics, economics, and immigration dissuades reporters from aggressively covering these beats or even covering them at all. At the pipeline level, violence leads reporters to leave or consider leaving the profession entirely.⁶⁸Carroll, supra note 65, at 411.

Illustrating these dynamics more specifically, one study found that 37% of female journalists who reported that they had been threatened, harassed, or attacked indicated they avoided certain stories; 8% changed the content or perspectives of the news that they write about; 16% considered requesting a transfer or different beat and 6% requested a transfer or different beat; and 29% indicated the threats and attacks they received made them think about getting out of the profession.⁶⁹Michelle Ferrier, TrollBusters & Int’l Women’s Media Found., Attacks and Harassment: The Impact on Female Journalists and Their Reporting 39, 44 (Elisa Lees Munoz ed., 2018), https://www.iwmf.org/wp-content/uploads/2018/09/Attacks-and-Harassment.pdf [https://perma.cc/L4ZU-BCE9].

As additional examples, threats against election officials deter those officials from doing work necessary to the democratic self-governance at the core of the First Amendment. A nationwide study of local election officials found that more than one in three of those surveyed reported that they have experienced threats, harassment, or abuse because of their work; more than half reported that they worry about their colleagues’ safety; more than a quarter worried about their own physical safety; and more than a third reported that they knew of local election officials who had left their jobs at least in part because of fear for their safety.⁷⁰The Brennan Ctr. for Justice, Local Election Officials Survey 9, 15, 19 (2024), https://www.brennancenter.org/our-work/research-reports/local-election-officials-survey-may-2024 [https://perma.cc/C3EN-WUDH]. Not surprisingly, many local jurisdictions now report high levels of turnover among their election officials.⁷¹Miles Park, In Some States, More than Half of the Local Election Officials Have Left Since 2020, NPR (Sept. 26, 2023, 5:12 PM), https://www.npr.org/2023/09/26/1200616113/election-official-threats-harassment-turnover [https://perma.cc/7T97-FKL2]; Michael Beckel, Amelia Minkin, Amisa Ratliff, Ariana Rojas, Kathryn Thomas & Adrien Van Voorhis, Issue One, The High Cost of High Turnover 1 (2023), https://issueone.org/wp-content/uploads/2023/09/The-High-Cost-of-High-Turnover-Report.pdf [https://perma.cc/M53Z-NKD3].

II. Attending to Targets’ Free Speech Interests in Threats and Stalking Cases

As Frederick Schauer observed, courts often choose to err on the side of overprotecting defendants’ speech when crafting First Amendment doctrine:

The chilling effect doctrine reflects the view that the harm caused by the chilling of free speech (or other protected activity) is comparatively greater than the harm resulting from the chilling of the other activities involved. And, the logical and necessary mandate of the chilling effect doctrine is that legal rules be formulated so as to allocate the risk of error away from the preferred value, thereby minimizing the occurrence of those errors which we deem the most harmful.⁷²Schauer, supra note 10, at 705.

The Counterman majority made precisely this choice when it required proof of a speaker’s recklessness as predicate to identifying a “true threat” unprotected by the First Amendment.⁷³See supra notes 15–17, 28–31 and accompanying text.

Again, that is a choice—but by no means the only justifiable, and speech-protective, choice. Because speech interests lie on both sides of threats and stalking cases, those cases require courts to choose not only whether to prefer speakers’ expression over incommensurate harms experienced by targets (like harms to their mental and physical health and their quality of life) but also whether to protect the speaker’s expression at the expense of the target’s expression. This Part considers possibilities for treating targets’ and speakers’ expressive interests in threats and stalking cases with at least the same regard.

A. The First Amendment Argument for Symmetrical Concern for Speakers’ and Targets’ Expressive Interests in Threats Cases

Neither the Counterman majority nor Billy Counterman’s own briefing offered any evidence that anyone’s speech—much less Counterman’s—had been chilled by an objective listener-centered approach (like Colorado’s) for identifying unprotected true threats.⁷⁴See Brief on the Merits for Respondent at 38–39, 45, Counterman v. Colorado, 143 S. Ct. 2106 (2023) (No. 22-138) (pointing out that Counterman’s anecdotal discussion of fact patterns offered to suggest an objective standard’s potential for chilling did not involve speech prosecuted under a threats theory or involved speech prosecuted under Counterman’s proposed specific intent standard). As dissenting Justice Barrett observed, “objective tests are effectively the status quo today, yet Counterman still struggles to identify past prosecutions that came close to infringing on protected speech.”⁷⁵Counterman v. Colorado, 143 S. Ct. 2106, 2138 (2023) (Barrett, J., dissenting) (“Before we took this case, the vast majority of Courts of Appeals and state high courts had upheld [statutes that required only an objective showing] as constitutional.”). A number of scholars have noted the limited empirical support for chilling effects concerns. See Suneal Bedi, The Myth of the Chilling Effect, 35 Harv. J.L. & Tech. 267, 307 (2021) (questioning courts’ reliance on chilling effect concerns when crafting doctrine); Jennifer M. Kinsley, Chill, 48 Loy. U. Chi. L.J. 253, 253 (2016) (criticizing chilling effects concerns as making “too may false assumptions about the speakers’ knowledge of the law, their ability to correctly apply the law, and their willingness to conform to the law”); Kendrick, supra note 23, at 1675; Penney, supra note 62, at 1454–55, 1470 (challenging conventional understandings of the chilling effect as “empirically weak” and asserting that most people “are often not sufficiently aware of the law or state activities such that any possible legal harm or sanction could impact their decision about speaking or acting.”); Schauer, supra note 10, at 730 (“While the chilling effect concept appears to be premised upon predictions or assumptions about human behavior, no evidence has been proffered to justify those predictions. It has not been clearly established that individuals are mistakenly deterred or become overly cautious as a result of the existence of particular statutes, rules, or regulations.”). Nevertheless, concern about the law’s chilling effect—that is, the “overdeterrence of benign conduct that occurs incidentally to a law’s legitimate purpose or scope”⁷⁶Kendrick, supra note 23, at 1649 (emphasis omitted).—is legitimate, indeed intuitively powerful, as a background assumption about how the world works.⁷⁷See Joseph Blocher & Reva B. Siegel, When Guns Threaten the Public Sphere: A New Account of Public Safety Regulation Under Heller, 116 Nw. U. L. Rev. 139, 186–89 (2021) (discussing the background assumptions that inform the Court’s constitutional doctrine in various areas despite the lack of empirical evidence for such assumptions). For this reason, the Counterman majority was not wrong to worry about chilling effects even absent anecdotal or empirical evidence of chilling.

But courts committed to robust free speech protections should also credit the equally valid premise that a legal standard that fails adequately to protect the targets of threats and stalking will enable the silencing of targets’ speech. Indeed, the First Amendment case for attending to targets’ free speech interests is at least as strong as that for attending to chilling concerns given that targets’ fear for their physical safety demonstrably deters their expression.⁷⁸See supra notes 62–71 and accompanying text. Just as a speaker’s First Amendment defense need not include evidence that their—or anyone’s—speech was chilled by the prospect that they could be held liable for objectively terrifying statements, a prosecutor or plaintiff need not prove that the target’s—or anyone’s—speech was silenced by their reasonable fear for their physical safety. In short, when developing and applying First Amendment doctrine, courts should weigh the free speech costs borne by the targets of threats and stalking as heavily as the free speech costs borne by speakers chilled by the prospect of the government’s legal action.

The First Amendment argument for this approach gathers additional force when we recognize that targets of threats and stalking cannot meaningfully rely on the traditional self-help remedies of counterspeech and avoidance.⁷⁹See Whitney v. California, 274 U.S. 357, 377 (1927) (Brandeis, J., concurring) (“If there be time to expose through discussion the falsehood and fallacies, to avert the evil by the processes of education, the remedy to be applied is more speech, not enforced silence.”). For example, only 9% of stalking victims surveyed who tried to discourage their stalkers reported that their engagement improved the situation.⁸⁰Mary P. Brewster, An Exploration of the Experiences and Needs of Former Intimate Stalking Victims 47 (1998), https://www.ojp.gov/pdffiles1/nij/grants/175475.pdf [https://perma.cc/7VZM-94AX]; see also James Geistman, Brad Smith, Eric G. Lambert & Terry Cluse-Tolar, What to Do About Stalking: A Preliminary Study of How Stalking Victims Responded to Stalking and Their Perceptions of the Effectiveness of These Actions, 26 Crim. Just. Stud. 43, 60 (2013) (“Our findings suggest that stalking victims who confronted stalkers on their own, regardless of whether they were victims of violent or nonviolent stalking, were likely to perceive that their efforts were ineffective. In some cases, their responses had the opposite effect and the victims reported that the stalking behavior worsened.”). A target’s engagement with a stalker through counterspeech can aggravate and escalate the stalker’s behavior⁸¹See Mullen & Pathé, supra note 59, at 294 (reporting that some “stalkers will react with extreme violence to their victim’s repeated rebuffs”). and can fuel a stalker’s delusions of an actual relationship with the target by “gratify[ing] the stalker’s wishes to have, and to hold onto, a relationship and reinforc[ing] the pursuit.”⁸²Id.; see also id. at 310 (reporting that “any contact with the perpetrator, however intermittent, will reinforce the unwanted behavior”); Mary Anne Franks, How Stalking Became Free Speech: Counterman v. Colorado and the Supreme Court’s Continuing War on Women, Geo. Wash. L. Rev. On Docket (2022), https://www.gwlr.org/how-stalking-became-free-speech-counterman-v-colorado-and-the-supreme-courts-continuing-war-on-women [https://perma.cc/E2BZ-MYZK] (“[S]talkers often sincerely believe that their behavior is welcome. These delusional beliefs make them more, not less, dangerous to their victims.” (footnote omitted)). For these reasons, targets are often counseled not to engage with stalkers.⁸³See Stalking Prevention, Awareness & Res. Ctr., Stalking Safety Strategies 2 (2022), https://www.stalkingawareness.org/wp-content/uploads/2022/05/Safety-Strategies.pdf [https://perma.cc/FP8G-KMH2] (“Consider cutting off any and all communication with the stalker. Many stalkers misinterpret any contact (even negative contact) as encouragement.”). Nor, due to limited resources, can many targets protect themselves through avoidance. “[M]oving to a new home or changing jobs or schools to avoid threatened violence are less likely to be options for domestic violence victims than for other threat victims,” legal scholar Jessica Miles explains.⁸⁴Miles, supra note 50, at 736; see also id. at 736–37 (“The relationship between poverty and the increased likelihood of violence is further supported by research that shows that domestic violence victims with the fewest resources experience the highest rates of repeat abuse.”).

A target-centered approach is not without precedent, as the Supreme Court sometimes chooses to shape categories of less-protected speech in ways that privilege listeners’ interests when those listeners cannot protect themselves from harmful speech through rebuttal or escape: this is often the case when speakers enjoy advantages of information or power (or both) over their listeners.⁸⁵Elsewhere, I have examined settings involving these asymmetries in more detail. See generally Helen Norton, What Twenty-First-Century Free Speech Law Means for Securities Regulation, 99 Notre Dame L. Rev. 97 (2023) (discussing corporations’ speech to investors and shareholders); Helen Norton, Manipulation and the First Amendment, 30 Wm. & Mary Bill Rts. J. 221 (2021) (discussing online platforms’ interactions with their users); Helen Norton, Discrimination, the Speech That Enables It, and the First Amendment, 2020 U. Chi. Legal F. 209 (2020) (discussing employers’ speech to workers); Helen Norton, Powerful Speakers and Their Listeners, 90 U. Colo. L. Rev. 441 (2019) [hereinafter Norton, Powerful Speakers and Their Listeners] (describing asymmetries of information and power between speakers and listeners in several contexts); Helen Norton, Pregnancy and the First Amendment, 87 Fordham L. Rev. 2417 (2019) (discussing the speech of those providing reproductive health care services to pregnant women); Helen Norton, Robotic Speakers and Human Listeners, 41 Seattle U. L. Rev. 1145 (2018) (discussing communications produced by artificial intelligence); Helen Norton, Truth and Lies in the Workplace: Employer Speech and the First Amendment, 101 Minn. L. Rev. 31 (2016) (discussing employers’ speech to workers); Helen Norton, Secrets, Lies, and Disclosures, 27 J.L. & Pol. 641 (2012) (discussing campaign speakers’ and donors’ speech to voters). Recall the Court’s defamation jurisprudence, which requires public officials and public figures to show a speaker’s recklessness (in other words, “actual malice”) before imposing liability for that speaker’s reputation-damaging falsehoods, while relaxing the showing required of private-figure plaintiffs with less ability to remedy reputational harm themselves through counterspeech. More specifically, the Court enables private figures to recover presumed and punitive damages for defamatory falsehoods on matters of private concern without any showing of the speaker’s recklessness.⁸⁶Dun & Bradstreet, Inc. v. Greenmoss Builders, Inc., 472 U.S. 749, 761 (1985) (plurality opinion). The Supreme Court also understands the First Amendment to permit defamation plaintiffs who are private figures to recover actual damages for defamatory falsehoods on matters of public concern upon a showing of the speaker’s negligence (but requires proof of the speaker’s actual malice to recover presumed or punitive damages in such cases). Gertz v. Robert Welch, Inc., 418 U.S. 323, 347–50 (1974). While this more target-friendly choice rests in great part on the lower First Amendment value of reputation-damaging falsehoods that do not address matters of public concern, it also turns on private figures’ more limited ability (compared to public officials) to protect their reputational interests through rebuttal:

Public officials and public figures usually enjoy significantly greater access to the channels of effective communication and hence have a more realistic opportunity to counteract false statements than private individuals normally enjoy. Private individuals are therefore more vulnerable to injury, and the state interest in protecting them is correspondingly greater.⁸⁷Gertz, 418 U.S. at 344 (footnote omitted).

When balancing speakers’ expressive interests against private figures’ non-speech (that is, reputational) interests for defamation purposes, the Court thus requires targets to prove no more than a speaker’s recklessness—and sometimes only their negligence. More powerful still is the First Amendment argument for privileging targets’ free speech interests over those of speakers’, especially when those targets cannot meaningfully protect themselves through the traditional self-help remedies of exit and voice. In other words, although recklessness might be an appropriate First Amendment compromise in threats cases if courts were simply balancing defendants’ speech interests against targets’ incommensurate non-speech interests in health and quality of life, recklessness is a much less obvious First Amendment choice if we attend to both parties’ speech interests.

As an illustration, consider the Court’s listener-centered understanding of the First Amendment that treats “false or misleading” commercial speech as entirely unprotected by the First Amendment because of the harm it inflicts on targets’ (consumers’) First Amendment interests in making informed and autonomous decisions regardless of the commercial speaker’s mental state.⁸⁸Cent. Hudson Gas & Elec. Corp. v. Pub. Serv. Comm’n of New York, 447 U.S. 557, 563–64 (1980) (“The First Amendment’s concern for commercial speech is based on the informational function of advertising. Consequently, there can be no constitutional objection to the suppression of commercial messages that do not accurately inform the public about lawful activity.” (citation omitted)); see also Zauderer v. Off. of Disciplinary Couns., 471 U.S. 626, 638 (1985) (“The States and the Federal Government are free to prevent the dissemination of commercial speech that is false, deceptive, or misleading . . . .”). This listener-centered approach also recognizes that consumers cannot meaningfully protect themselves from these informational harms, because commercial actors have considerably greater access to accurate information about their own goods and services than do consumers.⁸⁹See Norton, Powerful Speakers and Their Listeners, supra note 85, at 446–48 (explaining commercial actors’ informational advantages over consumers). For these reasons, many false advertising laws have long prohibited false or misleading commercial speech—regardless of the commercial speaker’s mental state—because of the harm posed to consumers as the targets of such speech.⁹⁰E.g., Aaron v. SEC, 446 U.S. 680, 696–702 (1980) (Section 17(a)(3) of the Securities Act of 1933 “quite plainly focuses upon the effect of particular conduct on members of the investing public, rather than upon the culpability of the person responsible”); FTC v. Algoma Lumber Co., 291 U.S. 67, 81 (1934) (explaining that advertisers’ lack of culpable mental state does not insulate them from liability under the Federal Trade Commission Act’s bar on deceptive trade practices). 

Because fear—and speech—are on both sides of threats and stalking cases, a robust understanding of the Free Speech Clause supports a standard similarly attentive to targets’ expressive interests. To be clear, my point is not that the First Amendment necessarily requires a listener-centered standard when the free speech interests of speakers and their targets collide. But neither does the First Amendment require a speaker-centered standard when we must choose between speakers’ and targets’ expression. I suggest instead that the First Amendment permits—indeed, supports—governmental choices to privilege targets’ free speech interests in certain settings. To this end, I urge that the First Amendment doctrine of threats and stalking cases treat the free speech costs borne by speakers and targets with at least equal concern. This symmetrical concern for speakers’ and targets’ expression in threats cases recognizes that objective listener-centered standards advance free speech as least as much as subjective speaker-centered standards. This approach thus understands the First Amendment to permit legislatures or courts to choose from a range of speech-protective standards such that the Counterman Court was wrong to deny Colorado’s Supreme Court that choice.

What might this symmetrical concern look like in practice? The rest of this Article identifies some possibilities.

B. The Counterman Court’s Missed Opportunity: Defining Unprotected True Threats Through an Objective Target-Centered Standard

Start with the road not taken by the Counterman majority: the Colorado Supreme Court’s objective target-centered approach for identifying unprotected true threats. Developed in a case involving high school students’ heated late-night Twitter (now known as “X”) argument shortly after a local school shooting, that context-driven standard required courts to consider the surrounding circumstances to determine whether “an intended or foreseeable recipient would reasonably perceive [the statement(s)] as a serious expression of intent to commit an act of unlawful violence.”⁹¹In re R.D., 464 P.3d 717, 721, 731 (Colo. 2020), abrogated by Counterman v. Colorado, 600 U.S. 66 (2023). In my view, this test treated speakers’ and targets’ free speech interests with equal concern by requiring courts to consider multiple factors that attend to both sets of interests.⁹²This Article focuses on the appropriate First Amendment analysis for threats and stalking cases that involve demonstrably direct collisions between speakers’ and targets’ speech interests—and to be sure, this includes attention to the speech-silencing harm of hate speech that takes the form of threats or stalking. See Virginia v. Black, 538 U.S. 343, 363 (2003) (considering cross-burning’s potential as a “particularly virulent” type of threat given its “long and pernicious history as a signal of impending violence”). This Article does not, however, address the First Amendment analysis to be applied to hate speech or harassment that does not cause the target to fear for their physical safety. See Snyder v. Phelps, 562 U.S. 443, 458–59 (2011) (holding that the First Amendment protected hate speech on a matter of public concern in a context that did not cause a target’s potentially speech-silencing fear for their physical safety). Note, however, that objective standards like Colorado’s approach also protect the equality interests of members of marginalized communities who both disproportionately experience enforcement action as speakers and disproportionately experience threats and stalking as targets. See Counterman, 143 S. Ct. at 2122–23 (Sotomayor, J., concurring in part and concurring in the judgment) (“The burdens of overcriminalization will fall hardest on certain groups . . . . Members of certain groups, including religious and cultural minorities, can also use language that is more susceptible to being misinterpreted by outsiders. And unfortunately, yet predictably, racial and cultural stereotypes can also influence whether speech is perceived as dangerous.”); Danielle Keats Citron, Cyber Civil Rights, 89 B.U. L. Rev. 61, 69–81 (2009) (describing how cyberharassment disproportionately targets women and people of color). These factors included:

• The statement’s role in any broader exchange or events—for example, whether the statement “was spontaneous or [instead] responsive to some other communication,” where statements that matched the “overall tone” of a conversation in which it joined were less likely to be experienced as objectively terrifying.⁹³In re R.D., 464 P.3d at 732.
• The medium or platform through which the statement was communicated (including any “distinctive conventions or architectural features”)—where, for example, “prevailing norms in a particular genre” or forum might “recast violent language in a less threatening light.”⁹⁴Id. at 731–32; see also Lyrissa Barnett Lidsky & Linda Riedemann Norbut, #I🔫U: Considering the Context of Online Threats, 106 Calif. L. Rev. 1885, 1910–28 (2018) (discussing how platforms’ architectural features affect the meaning of posts on those platforms and the value of expert witnesses in explaining the meaning of speech in these settings).
• The manner in which the statement was conveyed (“e.g., anonymously or not, privately or publicly”⁹⁵In re R.D., 464 P.3d at 722.)—in which, depending on the context, the speaker’s anonymity and the choice to personally target the statements rather than direct them to a large public audience might cause a reasonable target to experience the statements as threatening.
• Any relationship between the speaker and target—including any history of violence, the speaker’s awareness of the target’s particular vulnerabilities, or the target’s awareness of the speaker’s patterns of speech or emotional state.⁹⁶Id. at 733.
• The audience’s subjective reaction: Contrast Watts v. United States, in which the Court relied in part on the audience’s laughter to identify a statement as protected political hyperbole rather than an unprotected threat⁹⁷Watts v. United States, 394 U.S. 705, 707 (1969) (per curiam).with statements that prompt their audience to report their concerns about potential violence with law enforcement or trusted adults.⁹⁸In re R.D., 464 P.3d at 733.

C. Target-Attentive Approaches to Threats Cases After Counterman

Unless and until the Court chooses to revisit its decision in Counterman, however, an objective context-driven standard is precluded in threats cases, which now require prosecutors and plaintiffs to establish a speaker’s recklessness before imposing criminal or civil punishment.⁹⁹See supra note 3 and accompanying text. That the same First Amendment standard applies to both criminal and civil law settings (like civil protection orders) only raises the expressive stakes of this requirement.¹⁰⁰See Counterman v. Colorado, 143 S. Ct. 2106, 2140 (2023) (Barrett, J., dissenting) (“[T]his case is about the scope of the First Amendment, not the interpretation of a criminal statute. Accordingly, the Court’s holding affects the civil consequences for true threats just as much as it restricts criminal liability.”); see also id. at 2140–41 (canvassing the consequences of a recklessness requirement for targets’ ability to secure civil protections from threats through restraining orders, civil enforcement statutes, and school discipline). Counterman and its limitations now present new choices among available approaches, some more attentive to targets’ free speech interests than others.

First, because Counterman establishes recklessness as a First Amendment floor and not a ceiling,¹⁰¹See id. at 2139 (“The optimal balance strikes me as a question best left to the legislature, which could calibrate the mens rea to the circumstance—for example, higher for the criminal context and lower for the civil.”). legislatures remain free to impose even higher levels of mens rea (like intent or knowledge) as a condition of criminal or civil liability. And some do.¹⁰²E.g., Cal. Penal Code § 76 (prohibiting threats of “any elected public official” made with the speaker’s “specific intent that the statement is to be taken as a threat”); Mass. Gen. Laws ch. 258E, § 1 (2025) (defining harassment (in other words, actionable threats) as “3 or more acts of willful and malicious conduct aimed at a specific person committed with the intent to cause fear, intimidation, abuse or damage to property”). But Counterman does not require legislatures to so choose. Attention to targets’ free speech interests thus supports the legislative choice, consistent with free speech values, not to require any subjective mental state greater than recklessness, which already leaves targets unprotected in the many threats cases in which targets lack access to evidence of the defendant’s interior mental state.¹⁰³See supra notes 4–5 and accompanying text.

Second, courts’ recklessness determinations in threats cases can sometimes be informed by targets’ experiences. The Counterman majority articulated its recklessness standard as focused on the defendant’s “insufficient concern with risk, rather than awareness of impending harm.”¹⁰⁴Counterman, 143 S. Ct. at 2117. A law requiring the defendant’s “awareness of impending harm,” in contrast, requires the prosecution to prove a defendant’s awareness “that [a] result is practically certain to follow.” Id. (quoting United States v. Bailey, 444 U.S. 394, 404 (1980)). As the Court explained in an earlier decision, “[t]hat risk need not come anywhere close to a likelihood. Speeding through a crowded area may count as reckless even though the motorist’s ‘chances of hitting anyone are far less [than] 50%.’ ”¹⁰⁵Borden v. United States, 141 S. Ct. 1817, 1824 (2021) (quoting Wayne R. LaFave, Substantive Criminal Law § 5.4(f) (3d ed. 2018)). To demonstrate a defendant’s conscious disregard “of a substantial risk” that his statements would cause his target to fear for her physical safety, prosecutors and plaintiffs may present circumstantial evidence of recklessness—that is, evidence that the defendant ignored “obvious” risks or knew of facts that would have made the danger obvious to someone in the defendant’s situation.¹⁰⁶See Farmer v. Brennan, 511 U.S. 825, 842 (1994) (Whether the defendant “had the requisite knowledge of a substantial risk is a question of fact subject to demonstration in the usual ways, including inference from circumstantial evidence, and a factfinder may conclude that [the defendant] knew of a substantial risk from the very fact that the risk was obvious.” (citation omitted)). To be sure, the defendant remains free to introduce evidence that they were unaware of risks that others would find obvious.¹⁰⁷See id. at 844 (“[I]t remains open to the [defendants] to prove that they were unaware even of an obvious risk to inmate health or safety. That a trier of fact may infer knowledge from the obvious, in other words, does not mean that it must do so.”). Whether these risks are obvious to the defendant can sometimes be informed by the target’s experience and behavior. For instance, a defendant’s awareness that his target had repeatedly blocked his calls, texts, or other messages—an awareness perhaps demonstrated by his creation of new accounts from which to message his target—can show that he was aware of “obvious” risks that his target experienced his communications as threatening. So, too, could his awareness that his target had changed her expressive (or other) behavior in response to his statements.¹⁰⁸See Lyrissa Barnett Lidsky & RonNell Andersen Jones, Of Reasonable Readers and Unreasonable Speakers: Libel Law in a Networked World, 23 Va. J. Soc. Pol’y & L. 155, 177 (2016) (predicting that skeptical juries “will rarely accept a defendant’s argument that she truly believed her delusional and defamatory statements”).

Of course, a target could ensure that a speaker was aware of this risk by telling him that she experienced his statements as threatening—but, as discussed above, such engagement can itself be existentially dangerous.¹⁰⁹See supra notes 82–83 and accompanying text. This remains among Counterman’s major deficits in choosing a recklessness inquiry focused on the defendant’s interior mental state rather than threats’ speech-silencing effects on their targets.

D. Target-Attentive Approaches to the Many Stalking Cases That Do Not Involve Threats and Where Counterman Thus Does Not Apply

As explained above, stalking laws protect targets from speech-silencing and life-disrupting fear in ways related to, but distinct from, the protections offered by threats laws.¹¹⁰See supra notes 38–41 and accompanying text. While threats laws address speech that causes its target to fear for their physical safety, stalking laws address fear induced by repeated and unwelcome conduct, contact, or communication. Many stalking cases thus do not, and need not, turn on any allegedly threatening content of the speaker’s communications. In those cases, the default First Amendment rules apply—not Counterman’s recklessness requirement.¹¹¹E.g., State v. Labbe, 314 A.3d 162, 179 (Me. 2024) (“Some stalking prosecutions, like Counterman’s, may rely in whole or in part on words used by a defendant to establish the ‘course of conduct’ and consequent effect upon the victim. It does not follow, however, that the Counterman standard applies to every stalking prosecution in which words are spoken or electronic communication devices are used. Rather, Counterman’s holding is clear: where the State relies on the content of a defendant’s expression as the basis for a stalking charge and to establish harm to the victim, the additional requirement to prove subjective mens rea of recklessness applies.”). Nevertheless, because Counterman addressed threats in the context of a stalking case, the danger remains that courts will inappropriately apply its recklessness standard to stalking cases that do not involve threats.¹¹²Indeed, for a time this was the case in Colorado, until the state supreme court overruled a trial court’s holding that a prosecutor had to prove recklessness in a stalking case even though the prosecutor made clear that he would not be proceeding under a true threats theory. People v. Crawford, No. 24SA226, slip op. at *P2 (Colo. May 12, 2025) (holding that “the charges the prosecution brought here, carefully based on repeated actions—including contacts (i.e., texts, phone calls, and emails) but not their contents— do not require proof that the defendant communicated or otherwise acted with a reckless state of mind.”). In the interest of full disclosure, I note that I served on the team representing Colorado’s Attorney General in an amicus filing before the state supreme court in Crawford.

So the first step in attending to stalking targets’ free speech interests is to make sure that courts understand that Counterman’s recklessness requirement does not apply to stalking cases that do not allege unprotected true threats.¹¹³See Citron, supra note 32, at 204 (“Counterman did not rule that unprotected true threats are necessary for cyber stalking convictions to comport with the First Amendment. But law enforcers could get the [mistaken] impression that now there must be proof of recklessly made threats in any cyber stalking case.”). The second step is to engage targets’ free speech interests when applying the appropriate First Amendment analysis in the many stalking cases that do not involve threats—and where the free speech costs to the target remain substantial while the defendant’s expressive interests are low (since the repeated and unwelcome nature of stalking, rather than its content, often induces fear).¹¹⁴See Lakier & Douek, supra note 2, at 192 (discussing how an endless of barrage of messages can cause fear regardless of their content: “Their mere presence in her inbox, the fact of their constant arrival, that they just did not stop, made [the defendant’s messages] distressing regardless of what they said”).

To illustrate how courts’ assessment of First Amendment challenges to the enforcement of anti-stalking laws can attend to targets’ free speech interests, consider the following fact patterns:

• Defendant (D) repeatedly follows and watches Target (T) as they walk to and from their car to their office, gym, grocery store, and home;
• D calls T every morning and hangs up as soon as they answer;
• D calls T every morning and says, “Hello, beautiful” when T answers; and
• D calls T every morning and says, “Die” when T answers.

Each of these fact patterns involves repeated and unwelcome behaviors that, depending on the circumstances, could cause a reasonable person to experience fear and thus violate a state’s stalking statute.¹¹⁵E.g., Colo. Rev. Stat. § 18-3-602(1)(c) (2022) (making it unlawful “[r]epeatedly [to] follow[], approach[], contact[], place[] under surveillance, or make[] any form of communication with another person” in “a manner that would cause a reasonable person to suffer serious emotional distress and does cause that person . . . to suffer serious emotional distress”). But some of the fact patterns do so in ways that involve no communication—or if they do involve communication, they do so regardless of the communication’s content.

In short, as this Section explains, the government’s prosecution of conduct normally triggers no First Amendment review. If a defendant nevertheless seeks to challenge the prosecution of conduct on First Amendment grounds, the defendant would have to establish that their conduct was sufficiently expressive to warrant First Amendment protection. If a court determines the conduct to be expressive, then intermediate scrutiny applies. If a defendant challenges the content-neutral prosecution of speech (for example, repeated and unwelcome communications regardless of any threatening content), then again intermediate scrutiny applies. These analyses can, and should, foreground targets’, as well as speakers’, free speech interests.

1. Stalking That Takes the Form of Conduct Only

Fact patterns A (repeatedly following and watching the target) and B (repeatedly dialing the target’s phone number and hanging up without speaking) involve stalking that takes the form of repeated unwelcome conduct, not speech, since the speaker makes no oral or written communication. Consider these additional illustrations of such courses of conduct: a speaker that “send[s] envelopes of unknown white powder to the victim in the mail;” “repeatedly infect[s] the victim’s computers with viruses;” “open[s] unwanted on-line dating profiles under the victim’s identity;” or “arrange[s] every day for deliveries to be made at the victim’s home at all hours of the night.”¹¹⁶United States v. Ackell, 907 F.3d 67, 73 (1st Cir. 2018) (discussing how stalking often takes the form of conduct rather than speech).

Here, the government seeks to regulate what the defendant did, not what they said.¹¹⁷See Rumsfeld v. F. for Acad. & Institutional Rts., Inc., 547 U.S. 47, 60 (2006) (explaining that federal law regulates “conduct, not speech” when “[i]t affects what law schools must do—afford equal access to military recruiters—not what they may or may not say”). (As Genevieve Lakier and Evelyn Douek explain, “the term stalking was borrowed from the hunting context to emphasize the similarities between the way a hunter stalks their prey and the practices of surveillance and following that characterized the first well-publicized stalking cases.”¹¹⁸Lakier & Douek, supra note 2, at 147.) Stalking laws thus frequently address conduct with only incidental burdens on expression—triggering (and generally satisfying) no more than intermediate scrutiny.

More specifically, the First Amendment analysis of the government’s regulation of such repeated and unwelcome conduct should either be rational basis review (if the court determines the conduct to be nonexpressive) or intermediate scrutiny (if the court determines the conduct to be expressive).¹¹⁹Rumsfeld, 547 U.S. at 65–66 (considering whether regulated conduct is sufficiently expressive to trigger First Amendment review). Determining whether regulated conduct is sufficiently expressive to trigger intermediate scrutiny generally turns on whether the actor intends to communicate a message through their conduct and whether onlookers reasonably understand the actor to be delivering a message through their conduct.¹²⁰Texas v. Johnson, 491 U.S. 397, 404 (1989); Spence v. Washington, 418 U.S. 405, 410–11 (1974) (per curiam).

When a speaker repeatedly follows and watches their target, do they intend to communicate a message, and does their target understand them to be communicating a message? If, depending on the facts, the answer is “no,” deferential rational basis review will apply to the government’s enforcement.¹²¹E.g., Corrigan v. State, No. A23-1942, 2024 Minn. App. Unpub. LEXIS 604 at *10 (Minn. Ct. App. July 22, 2024) (“[The challenger’s] conviction here was based on his conduct in following, monitoring, or pursuing the other driver and not on the content of his expressions or speech. Thus, the holding in Counterman does not apply to [his] case.”). Perhaps some will take the view, however, that this conduct is expressive. For example, a speaker who repeatedly calls and hangs up on their target without saying anything (maybe only breathing heavily) might seek to send, and might be understood as communicating, a message of possessiveness, hostility, or even love toward the target such that application of the stalking law regulates repeated and unwelcome conduct that incidentally burdens expression.¹²²See United States v. O’Brien, 391 U.S. 367 (1968) (applying intermediate scrutiny to the government’s restriction of expressive conduct and holding that the law satisfied such scrutiny because it was narrowly tailored to the government’s significant regulatory interest). Here, too, the First Amendment analysis should remain attentive to the target’s expressive interests. For example, the application of stalking law to this conduct satisfies intermediate scrutiny when the government’s regulatory interests are unrelated to the suppression of ideas and when the enforcement action instead seeks to address the life-disrupting—including speech-silencing—effects of conduct that has little if any First Amendment value of its own.¹²³See R.A.V. v. City of St. Paul, 505 U.S. 377, 390 (1992) (“Where the government does not target conduct on the basis of its expressive content, acts are not shielded from regulation merely because they express a discriminatory idea or philosophy.”); Wisconsin v. Mitchell, 508 U.S. 476, 487 (1993) (explaining that state hate crimes statute targets violent conduct unprotected by First Amendment).

The Supreme Court’s analysis of First Amendment challenges to the enforcement of hate crimes laws to punish certain acts of violence illustrates these dynamics. Even when a defendant’s violence might seek to communicate, and be understood as communicating, a message of hatred based on the target’s protected class status, the Court has described hate crimes laws as the government’s constitutionally permissible regulation of conduct.¹²⁴Mitchell, 508 U.S. at 487. Along the same lines, the Court has stated that the First Amendment poses no bar to the government’s regulation of verbal harassment in the workplace that “produce[s] a violation of Title VII’s general prohibition against sexual discrimination in employment practices.”¹²⁵R.A.V., 505 U.S. at 389–90. In these cases, the Court did not state whether the regulated conduct—violent hate crimes and workplace harassment—was sufficiently expressive to trigger intermediate scrutiny, suggesting instead that the conduct did not merit First Amendment protection regardless of the review applied.¹²⁶See id.; see also Mitchell, 508 U.S. at 487. Lower courts have taken similar approaches, for example, when holding that laws prohibiting the physical obstruction of reproductive health care facilities to ensure access to lawful health care services satisfy intermediate scrutiny even though the regulated obstruction may seek to communicate a message and be understood as communicating a message.¹²⁷Terry v. Reno, 101 F.3d 1412, 1419 (D.C. Cir. 1996).

2. Stalking That Takes the Form of Communications That Instill Fear Regardless of Their Content

Turn next to stalking prosecutions of communications that cause fear because of their repeated and unwelcome nature—in other words, because of those communications’ frequency, volume, or persistence rather than their content.¹²⁸See Counterman v. Colorado, 143 S. Ct. 2106, 86 (Sotomayor, J., concurring in part and concurring in the judgment) (“The content of the repeated communications can sometimes be irrelevant, such as persistently calling someone and hanging up, or a stream of ‘utterly prosaic’ communications.” (quoting Id. at 2111–12)); State v. Labbe, 314 A.3d 162, 179–80 (Me. 2024) (“The ‘course of conduct’ for which [the defendant] was indicted and convicted involved a series of electronic communications—phone calls and texts—to the victim during a period of several weeks. Viewed in the context of the record as a whole, the stalking charge here was not predicated on the content of those communications but rather on the act of communicating itself—the repeated, unwelcome contact carried out through electronic devices—even after he was asked to stop; even after his possessions were returned; and even after he had been served with a protection order prohibiting him from having any direct or indirect contact with the victim. . . . This is plainly evident from the record. The content of the calls and messages was not threatening (‘utterly prosaic’); some were devoid of meaningful content; some were unanswered or were merely hang-ups; one consisted of dead air and just breathing.”). This can be the case, for example, of fact pattern C, where the defendant called the target every day with prosaic messages like “Hello beautiful.” Or by leaving a daily message on the target’s answering machine playing the target’s favorite song. Along the same lines, consider a speaker who made:

“multiple [phone calls] throughout the evening, often ‘back-to-back,’ ” and “into the early morning hours,” despite being told that he “needed to stop calling” and “I don’t want you calling this house anymore.” When the complainant unplugged her phone, the defendant then called each of her parents and he was again told not to call. A few days later, he showed up at the complainant’s apartment building and “stood outside for a couple of minutes,” until he was directed to leave by security. Then he returned the following day again, and “security informed him that he was not permitted in the building.” Several months later, he again placed “nonstop calls” to the complainant from 1:30 a.m. to 3:00 a.m., and during that time he again showed up to the complainant’s apartment and tried calling her from the apartment’s call box. He then tried sneaking into the apartment building to get to the complainant, but was again thwarted. Notice that we have not said one word about the content of the defendant’s speech . . . because it was immaterial to his stalking conviction. He could have been trying to sell his victim a vacuum cleaner and the above actions would still have amounted to stalking.¹²⁹Mashaud v. Boone, 295 A.3d 1139, 1161 (D.C. 2023) (citations omitted) (describing the facts in Atkinson v. United States, 121 A.3d 780 (D.C. 2015)).

The content-neutral regulation of such repeated and unwelcome communications again triggers intermediate scrutiny, in which courts again assess whether the restriction is narrowly tailored to serve a significant government interest. Courts undertaking this intermediate scrutiny consider, among other things, whether the government’s content-neutral regulation leaves open ample alternative means of expression.¹³⁰See Frisby v. Schultz, 487 U.S. 474, 484–88 (1988) (upholding a town’s content-neutral ban on picketing targeted at specific homes as narrowly tailored to achieve the government’s significant interest in protecting residential privacy, especially because it left untouched ample alternative means of communication like distributing literature door-to-door). Moreover, the Supreme Court has made clear that the government’s content-neutral speech regulation “need not be the least restrictive or least intrusive means of” achieving its interests; instead, the government’s content-neutral regulation of speech satisfies intermediate scrutiny when it advances a substantial government interest that would otherwise be achieved less effectively.¹³¹Ward v. Rock Against Racism, 491 U.S. 781, 798–99 (1989).

Here, too, this analysis can and should foreground targets’—not just speakers’—free speech interests. The content-neutral application of stalking law to repeated and unwelcome communications regardless of their content satisfies intermediate scrutiny when the government’s regulatory interests are unrelated to the suppression of ideas and when the enforcement action instead seeks to address the life-disrupting—including speech-silencing—effects of repeated and unwelcome communications. That the traditional self-help remedies of exit and voice are so often ineffective—and indeed, frequently dangerous—for their targets further support the enforcement of stalking law as a narrowly tailored response.

3. Stalking That Takes the Form of Threatening Speech

Depending on the circumstances, fact pattern D (daily messages telling the target to “die”) might support a content-neutral stalking prosecution based on the communications’ repeated and unwelcome time, place, and manner, regardless of their content. And depending on the circumstances, fact pattern D may also support prosecution as true threats, which now requires proof of the defendant’s conscious disregard of the risk that his statements were causing his target to fear for her physical safety in addition to proof of each of the statutory elements. Note that these two possibilities require prosecutors and plaintiffs, as a strategic matter, to choose between distinct theories that trigger different First Amendment analyses.¹³²See Lakier & Douek, supra note 2, at 203 (“[L]ike most stalking laws, the Colorado law under which Counterman was convicted not only requires that the emotionally distressing communications be made on multiple occasions but also requires proof that those speech acts have a significant emotionally distressing effect on their recipient, and reasonably so. All of these requirements cabin the reach of the law and ensure that what it punishes is sustained and abusive communication, not an isolated, and therefore potentially misconstrued, speech act.”).

4. Stalking That Involves Both Conduct and Speech

The foregoing examples illustrate how the variety of repeated and unwelcome behaviors that constitute stalking can invite enforcement under different theories that trigger distinct First Amendment analyses. Some fact patterns involve repeated and unwelcome fear-inducing conduct—and, depending on the facts, that conduct may or may not be expressive. Some involve repeated and unwelcome speech that induces fear not because of its content but because of its repeated and unwelcome time, place, or manner. Some involve repeated and unwelcome speech that induces fear at least in part because of its threatening content.¹³³See, e.g., State v. Lindell, 828 N.W.2d 1, 2–3 (Iowa 2013) (upholding stalking conviction based on repeated contact that included both speech and conduct like a “handwritten note and flowers,” “hang-up calls,” physical surveillance, and damage to personal property). And some fact patterns may involve combinations that invite prosecutors and plaintiffs to make strategic choices among multiple theories.¹³⁴Again, theories that do not rely on evidence of speech of a threatening nature as an element of the case do not trigger Counterman’s recklessness requirement. But see Lakier & Douek, supra note 2, at 169 (expressing skepticism that stalking laws can be applied to communications in content-neutral ways, and canvassing the relevant regulatory tradition to suggest the possibility of a category of less-protected “unwanted, persistent and fear-producing speech” in which some showing of mens rea like recklessness might still be required). Whatever the First Amendment analysis, however, it should include attention to targets’, as well as defendants’, expressive interests.

Conclusion

Speakers who fear the prospect of the government’s punishment often express themselves less—or express themselves differently—than they would if they were not fearful. This, in turn, harms the speaker’s and the public’s First Amendment interests by stifling the delivery of opinions, facts, and ideas.

At the same time, targets fearful for their physical safety at the hands of a speaker often speak less, and speak differently, than they would if they were not fearful. And this, too, harms both the target’s and the public’s First Amendment interests.

Fear, and speech, are thus on both sides of threats cases and also many stalking cases. Courts’ doctrinal choices in threats and stalking cases sometimes chill speakers’ expression. And sometimes they instead permit the silencing of targets’ speech. A robust commitment to free speech, in my view, requires attention not only to the speech deterred by legal standards that are too quick to punish speakers but also to the speech silenced when the law fails adequately to protect the targets of threats and stalking from reasonable fear for their safety.

When courts ignore targets’ free speech interests to privilege defendants’ free speech interests, they pretend to make hard First Amendment problems easy. But principled problem-solving requires that we take care to recognize and explain what makes difficult problems difficult, as is the case here, where we must choose between speakers’ and targets’ important free speech interests.

Courts must, and do, make choices all the time when calibrating the categories of unprotected speech (for example, in true threats cases), and when otherwise applying First Amendment doctrine (for instance, in stalking cases that do not allege true threats). The Court’s choice in Counterman to privilege speakers’ free speech interests over targets’ was a choice—but that choice was neither inevitable nor necessarily speech-protective. When we illuminate, rather than obscure, the free speech injuries experienced by the targets of threats and stalking, we can identify doctrinal choices that attend to targets’ expressive interests as well as defendants’.